Wednesday, February 13, 2008

Brasyl: an appreciation

At TTACon 9 I had -- among many other discussions -- a talk with Paul Raven about Ian McDonald's novel Brasyl. I thought it was one of his best, Paul ranked it lower, and specifically lower than River of Gods. I then tried to make the point that -- for my money -- River of Gods was, metaphorically speaking, where McDonald's heart was (evidenced not only by its higher page count, but also by a slew of short stories set in the same world, released as a collection by Orion called Cyberabad Days), but that his mind was in Brasyl. And while there have been a veritable slew of raving reviews, there were also a couple that were not convinced of the novel's intellectual merit:

The big revelation at the end of the novel is perhaps not a hoary old cliche yet, but neither is it as fresh or strange as I would've expected from McDonald.

The real problem with Brasyl lies in the strange lack of invention that it brings, late on, to a fascinatingly weird premise. [...] when it comes to the crux of the book, the reader is surprised to find all the dark narrative mutterings about quantum theory and the multi–verse boiling down, after some fairly obvious foreshadowing, to a a thriller dénouement that involves a bunch of dimension-hopping assassins equipped with diminutive lightsabres.

The considerable amount of beer consumed that day most probably didn't make my argument very cogent, so I'll reformulate it -- hopefully in clearer terms -- here. For the record, I think McDonald delivered for a good 99,5%, and could have for the full 100% if he had given it a final twist, which ties in with Tim Martin's remarks. I'll get to that.

The high concept behind the novel is the existence of a simulated multiverse, which is generated by a Universe-wide quantum computer in the original, 'real' Universe. McDonald acknowledges the modern-day godfather of the multiverse David Deutsch (and specifically his thought-provoking The Fabric of Reality), and the idea of multiple realities being simulated on a Universe-wide computer stems from Frank Tipler's The Physics of Immortality (whose concept of an 'Omega Point' is also treated in The Fabric of Reality).

McDonald, however, does give his own twist to these concepts. For one, Tipler's 'Omega Point' -- where exponential growth of intelligence eventually enables control over the Universe over the largest possible scale -- was (at the time, 1994) believed to be taking place during the Big Crunch (Paul McAuley's Confluence trilogy takes Tipler's 'Omega Point' seriously). In the meantime, though, astronomers have found that the expansion of the Universe is accelerating (hence the search for the 'dark energy' that powers that acceleration), making a Big Crunch unlikely. So McDonald projects this 'Omega Point' on a forever-expanding Universe, where the multiversal simulation still is crunched, albeit at an ever-slower rate (which the inhabitants of the simulation won't notice, at least for a while, as their 'time' is simulated, as well).

The main point is, though, that the multiverse of Brasyl is a simulation. In The Fabric of Reality David Deutsch posits that the multiverse is real, and one of the main points of his arguments for that is the double slit experiment. This is one of the greatest puzzles of modern physics: the interference pattern also appears when the particles (photons or electrons) are sent through the double slit one by one, seemingly suggesting that one particle interferes with itself.

David Deutsch's solution to this mind-boggling phenomenon is that this single particle interferes with its ghost from another Universe, hence a reality that consists of more than one Universe. (Apart from the Copenhagen interpretation other scientists have tried to explain this like, for example, the late David Bohm (with his interpretation of the 'hidden variables' interpretation, called 'undivided whole' or 'ontological interpretation') and John Cramer -- with his 'transactional interpretation' where waves travel both forward and backward in time, which might be supported by Julian Barbour's view in The End of Time that [as far as I can understand] time is an emergent phenomenon, not a basic property, and thus time arises from the quantum background like temperature arising from the average kinetic energy of atoms: see also New Scientist issue 2639 of January 18, 2008; main article "The Time Illusion". But I digress...;-)

This ties in with the simulation argument: the notion that our reality might be nothing more than a very advanced simulation. This is an interesting thought experiment: *if* reality as we experience it is a simulation, then *how* would we know this?

The simple answer is: if the simulation is perfect, we would never know (and David Deutsch argues that a universal quantum computer can render an environment to any degree of accuracy. Well, maybe if that environment is limited. But not if that environment is a whole Universe, let alone a multiverse, I strongly suspect). There is no point, at any level, where this simulation would break apart, and for us -- the inhabitants of that simulation -- the difference between the perfect simulation and an *actual* reality cannot -- by definition -- be noticed (otherwise the simulation isn't perfect). For us it would literally make no difference (up until the point where the experimenters would stop the simulation).

Now, in my day job I'm a technical specialist. I've worked with a variety of high tech equipment, and in my experience a perfect machine does not exist. Extremely sturdy, immensely reliable machines: yes. But even these will never work perfect *all the time*. There's a name for perfect machines: perpetuum mobiles.

Therefore, I personally think that a perfect simulation is also impossible. If our reality is a simulation, there will be a point where the seams of this simulation will show. Reading Brasyl, the thought struck me that if indeed our reality is an extremely sophisticated simulation, maybe the simulation breaks apart at the quantum level, for instance with the crazy results of the double slit experiment?

Now, as far as I understand, this is not what McDonald is actually proposing in Brasyl: but he does posit that the three strands of his novel take place in parallel worlds of a simulated multiverse, and throughout the novel there are moments when the simulation shows through. Most clearly through the quantum knives(*), also through the golden frogs that, through observing a single photon, experience the whole multiverse, and I do remember a scene where Edson mentions that he sees a plane flying over Sao Paolo, very low, without hearing the engines (it's been two weeks since I read the book, and my memory certainly isn't perfect). As the novel progresses, these seams in the simulation become gradually more apparent (Marcelina's double, Fria's double, Father Luis's visions of the multiverse, etc.).

(*) = I strongly suspect that an *actual* knife that can cut through anything at the quantum level is impossible, because of the Uncertainty Principle. Cutting apart bonds between molecules or atoms: yes. Stripping all electrons from an atom, or severing nuclear components (protons & neutrons; or even quarks): I don't think so. However, the quantum knife in Brasyl operates in a simulated environment, so is the equivalent of a carefully aimed delete button, or a highly fine-tuned decompiling tool.

Ultimately, one might argue that setting the whole novel in an extremely sophisticated simulation is a copout: since it's all virtual, things can take place that are not possible in an *actual* reality. Like golden frogs seeing the multiverse through the observation of a single photon or a 'Q knife' that literally cuts through everything.

However, in my opinion McDonald played fair: he carefully foreshadowed this, very subtly at first, and increasingly as the novel progresses. Also, the liberties he takes are all necessary to bring the concept forward, and carefully measured out. He could have easily taken this freedom from physical laws to excess, but he didn't: restraint is, in this case, the mother of invention.

More importantly, the major players in Brasyl -- and the reader -- finally come to grips with the (simulated) world they live in. So what do they do with that knowledge? As mentioned above, Tim Martin was disappointed that it all only led to a thriller game of assassins fighting each other.

I remember someone commenting (not sure of this was from some review or blog post I read, or from an actual conversation: forgive my fleeting memory) that having a (French) inventor travel down the Amazon in 1732 with the equivalent of an analytical engine, some 60 years before Charles Babbage was born, thus enabling Dr Falcon to understand, way ahead of his time, the real implications of his situation, was too much of a sleight of hand.

Against this I would argue that somebody must find out the actual 'fabric of (un) reality' first, and that in an extremely sophisticated multiverse simulation this will be at the earliest possible point in history (the most unlikely, yet still possible early discovery of all the possible reality simulations).

And that is what the Father Luís/Dr Falcon/Diego Gonçalves thread is about: this is where the seams of the simulation came apart first. This where the larger implications of the situation were interpreted: and since they had 18th Century mindsets the larger picture they drew was predominantly religious: Diego Gonçalves believes that the simulation must not be disturbed as it is God's plan; while 'the Order': the Father Luís/Dr Falcon group eventually (not immediately in the very beginning, but increasingly as they explore their multiverse) realise that the Mother Universe on which their simulation is running is running out: the accelerated expansion is slowly ripping it apart, and this will ultimately disrupt all the simulated worlds. So 'the Order' wants to break out from the simulation into the *actual* world, and the Gonçalves group believes this is blasphemy of the highest degree, and sends out assassins to prevent this.

Admittedly, this does look very much like a thriller denouement: powerful cabals plotting each other's demise. However, they both think they are fighting for the highest stakes: to preserve the Universe. Even if they disagree on how it should be achieved.

The only thing that I found slightly -- minimally -- disappointing was that the '1732' thread did not make it unambiguously clear that *this* Father Luis/Dr Falcon world was the very, very first where the conflict between the Order and the Gonçalves group originated: it would have tied everything together, very neatly.

Possibly that's why McDonald didn't do it: this conflict would have started in more worlds than one, rendering this simulated multiverse a bit muddy, somewhat fuzzy. Almost like the real world...

Back in August 6, 2005 I sat next to David Pringle at the Hugo Award ceremonies. When the Hugo for best novel was presented, David told me that he was rooting for River of Gods, and I completely agreed with him. It lost to Jonathan Strange and Mr. Norrell --whose promotion budget and sales were an order of magnitude greater than that of River of Gods -- but by a relatively small margin. This year I hope Brasyl not only gets nominated, but that it gets the nod in Denver on August 9.